vocaloidfandomcom-20200222-history
Talk:Synthesizer V/@comment-153.107.192.207-20180831043158/@comment-53539-20180831101656
Bare in mind if English Vocals are done right given they are 5x bigger then Japanese Vocals, due to the larger samples amount, there is more variation across the library. More variation creates more realism since there is less repetition of the same samples (in theory). A good English Vocaloid should be able to sound far more realistic due to this alone then a Japanese Voclaoid who has 1/5 their sample amount in a single voicebank. Japanese vocaloids, however, have the advantage of precision because Japanese is a more precise language then English with exact sounds more or less and doesn't require odd traits like blending. In fact I think currently Japanese is the potentially least realistic as every other language has more samples, though I don't know enough of what the other 3 languages have to really say where they stand in the grand scheme of things. The same rule would apply across vocal synths in general when using ones that are sample based. And given most Vocaloids are Japanese, she would be able to sound realistic then most Japanese Vocaloids if circumstances are given right. I know a lot of people would disagree with me, but it turns out this is pretty much the case. I'm not going to go into this any more deeply though as this is an entire discussion of its own. However, even then if you download her and compare her vocals to another English vocal from Vocaloid, the Vocaloid is better overall in terms of realism, even if I compare it to my V2's... Given they aren't corrected English vocals and V2 is dated with no triphonesso they are held back by this... I still feel its not much difference at times and people are overhyping how realistic it sounds because of the demos. In the demos I noted that the vocals seem to have issues crossing large gaps of notes more so then Vocaloid and using the software... Don't go high to low or vice versa, its really ugly and my hunch was correct there. I mean Vocaloid is ugly too doing that, but this girl... She can't go more then 4 notes difference between each sound without it starting to have bad points being heard. But going from A-B-C, etc in a row is really quite smooth and pretty on par with Vocaloid. This synth has some great demos, but there is issues reported even from one or two of the demo makers that its not as good as it sounds. But bare in mind... UTAU is a clunky horrible engine at times and yet good examples of good producers using it would have you thinking otherwise and a lot of UTAU fans try to convince you their favourite vocal synth is not ugly because of this. It is, however, in my books after using it one of the best freebie software we have so I'm not complaining and given Yamaha has far greater resources, as does CeVIO's company I don't expect much from anyone else. It certainly should go far and is still quite young. It also doesn't have the capabilities that V5 has or CeVIO overall, so there are less options with this software when you use it compared to them. So there is better out there and there is more realistic out there, but for a free software this is one of the best options, if not the best, we've had in a long, long time. With UTAU out of the picture, its been for 3 years just CeVIO or Vocaloid for singing vocals, there has been no alternatives for those who can't use either. This has the Japanese fans excited for this reason. We had Alter/ego, but it was kinda bad as the engine that worked for Chipspeech wasn't good enough for realistic vocals. I will note despite the companys note you don't need multiple vocals with this software due to its extra function, it still isn't a full replacement for the need of extra vocals. There is more then just tension between multiple voicebanks and after using it I felt this is a claim that they aren't in a position that gives them a big advantage as their making it out to be. If you have vocalist who can voice act, they can do multiple very difference vocals not just based on tension but tone and other traits. A good example was Mel Blanc, who did most of the voices for the Looney Tunes and worked for several companies over his lifetime, there was a reason for his nickname "the man of a thousand voices". This software doesn't invalidate the need for extra vocals at all for this reason despite what it claims and this will become apparent if they ever make a vocal with more then one voice.